how to minimize the slippage?

Discussion in 'Order Execution' started by emk662, Apr 29, 2003.

  1. emk662


    I have a stock trading system that backtested very well. Now I am gonna put it into a real automated system and I am worried about the slippage, especially most of my trades are short period.
    How to minimize the slippage? Thanks.
  2. gms


    trade hi liquidity
    use limit orders
  3. You can't since it is out of your control so you must include it your backtest.

  4. What kind of style are you trading? Are you trading momentum, pairs, other? Best thing to do is to only use limit orders and don't chase a price down more than once. If you're trading nasdaq, it's not as critical since there's usually more liquidity on the big cap stocks, however, if you're trading NYSE, be careful because specialists can easily make the bid/ask spread 5-10 cents.

    If you are scalping momentum, NEVER EVER use market orders unless you want to be filled at the top or bottom of a strong move (esp on Nasdaq). However, an advanced strategy that I sometimes use when a stock is selling off like mad is I'd send in a market order of 100 shares to sell short. I'll have my buy 1100 shares ready loaded in my order bar and when I get filled on the 100 shares, I immediately press the buy 1100 shares at the ASK +0.01 or 0.02 to make sure I get the fill. Doesn't work on the up side because at the top, there's a downtick and you'll have a lot of trouble trying to short it without a bullet.

  5. I agree with HarryTrader, slippage should be included in your backtest (just as commissions should be). You have to be very conservative, a good system without factoring in slippage can turn into a bleeding system if you factor it in.
    Good Luck nonetheless.
  6. Just my 2 cents to save you some potential losses, I've spent a good amount of time researching this issue. I've backtested many many systems in TS for a few months, and found many high trade frequency systems that looked amazing in backtesting.

    In every case so far, adding realistic slippage figures killed these high frequency systems. Using slippage in your backtests is absolutely vital. A system that makes a lot of trades and great profits with no slippage figure added is probably worthless. If not, you may have a winner. Remeber, TS is applying your strategy to a last trade price, and doesn't account for the spread or slippage until you add it in. Last trade is not a realistic way to calc P/L, especially with equities.

    Also, always use 1 min look inside bar resolution. This will reveal other traps as well.

    Good luck
  7. emk662


    I trade pairs. I will probably send 300 trades once a day and each trades have $180,000 amount. Therefore, for a $20 stock, I trade 9000 shares; for a $40 stock, I trade 4500 shares. I am not sure to what extent this number of shares will affect the market and what the slippage will be. When we select pairs, we try to choose those with high daily amount (>$1.5million). And our pairs include stock in NASDAQ, NYSE and AMEX. By the way, what is the most liquid time in a day.

    Any ideas will be thanked.
  8. kowboy


    Trading Pairs?

    emk, could you briefly explain the basics of trading pairs for those of us that don't know. And how do you pick your candidates and are they usually in the same sector?

    Thank you
  9. emk662


    Yes, I trade two stocks from the same sector.